A hierarchical recursive method for text detection in natural scene images

• Published in: Multimedia tools and applications 2017-12, Vol.76 (24), p.26201-26223
• Main Authors: Wang, Xiaobing, Song, Yonghong, Zhang, Yuanlin, Xin, Jingmin
• Format: Article
• Language: English
• Subjects: Architecture; Change detection; Computer Communication Networks; Computer Science; Computer vision; Data Structures and Information Theory; Datasets; Image detection; Image segmentation; Multimedia Information Systems; Neural networks; Recursive methods; Special Purpose and Application-Based Systems; Texts
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-016-4099-2

Text detection in natural scene images is a challenging problem in computer vision. To robust detect various texts in complex scenes, a hierarchical recursive text detection method is proposed in this paper. Usually, texts in natural scenes are not alone and arranged into lines for easy reading. To find all possible text lines in an image, candidate text lines are obtained using text edge box and conventional neural network at first. Then, to accurately find out the true text lines in the image, these candidate text lines are analyzed in a hierarchical recursive architecture. For each of them, connected components segmentation and hierarchical random field based analysis are recursively employed until the detected text line no more changes. Now the detected text lines are output as the text detection result. Experiments on ICDAR 2003 dataset, ICDAR 2013 dataset and Street View Dataset show that the hierarchical recursive architecture can improve text detection performance and the proposed method achieves the state-of-art in scene text detection.